maintenance/cleanupImages.php

   1 <?php
   2 /**
   3  * Clean up broken, unparseable upload filenames.
   4  *
   5  * Usage: php cleanupImages.php [--fix]
   6  * Options:
   7  *   --fix  Actually clean up titles; otherwise just checks for them
   8  *
   9  * Copyright © 2005-2006 Brion Vibber <brion@pobox.com>
  10  * https://www.mediawiki.org/
  11  *
  12  * This program is free software; you can redistribute it and/or modify
  13  * it under the terms of the GNU General Public License as published by
  14  * the Free Software Foundation; either version 2 of the License, or
  15  * (at your option) any later version.
  16  *
  17  * This program is distributed in the hope that it will be useful,
  18  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  19  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  20  * GNU General Public License for more details.
  21  *
  22  * You should have received a copy of the GNU General Public License along
  23  * with this program; if not, write to the Free Software Foundation, Inc.,
  24  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  25  * http://www.gnu.org/copyleft/gpl.html
  26  *
  27  * @file
  28  * @author Brion Vibber <brion at pobox.com>
  29  * @ingroup Maintenance
  30  */
  31
  32 require_once __DIR__ . '/cleanupTable.inc';
  33
  34 /**
  35  * Maintenance script to clean up broken, unparseable upload filenames.
  36  *
  37  * @ingroup Maintenance
  38  */
  39 class ImageCleanup extends TableCleanup {
  40         protected $defaultParams = array(
  41                 'table' => 'image',
  42                 'conds' => array(),
  43                 'index' => 'img_name',
  44                 'callback' => 'processRow',
  45         );
  46
  47         public function __construct() {
  48                 parent::__construct();
  49                 $this->mDescription = "Script to clean up broken, unparseable upload filenames";
  50         }
  51
  52         protected function processRow( $row ) {
  53                 global $wgContLang;
  54
  55                 $source = $row->img_name;
  56                 if ( $source == '' ) {
  57                         // Ye olde empty rows. Just kill them.
  58                         $this->killRow( $source );
  59                         return $this->progress( 1 );
  60                 }
  61
  62                 $cleaned = $source;
  63
  64                 // About half of old bad image names have percent-codes
  65                 $cleaned = rawurldecode( $cleaned );
  66
  67                 // We also have some HTML entities there
  68                 $cleaned = Sanitizer::decodeCharReferences( $cleaned );
  69
  70                 // Some are old latin-1
  71                 $cleaned = $wgContLang->checkTitleEncoding( $cleaned );
  72
  73                 // Many of remainder look like non-normalized unicode
  74                 $cleaned = $wgContLang->normalize( $cleaned );
  75
  76                 $title = Title::makeTitleSafe( NS_FILE, $cleaned );
  77
  78                 if ( is_null( $title ) ) {
  79                         $this->output( "page $source ($cleaned) is illegal.\n" );
  80                         $safe = $this->buildSafeTitle( $cleaned );
  81                         if ( $safe === false ) {
  82                                 return $this->progress( 0 );
  83                         }
  84                         $this->pokeFile( $source, $safe );
  85                         return $this->progress( 1 );
  86                 }
  87
  88                 if ( $title->getDBkey() !== $source ) {
  89                         $munged = $title->getDBkey();
  90                         $this->output( "page $source ($munged) doesn't match self.\n" );
  91                         $this->pokeFile( $source, $munged );
  92                         return $this->progress( 1 );
  93                 }
  94
  95                 return $this->progress( 0 );
  96         }
  97
  98         /**
  99          * @param string $name
 100          */
 101         private function killRow( $name ) {
 102                 if ( $this->dryrun ) {
 103                         $this->output( "DRY RUN: would delete bogus row '$name'\n" );
 104                 } else {
 105                         $this->output( "deleting bogus row '$name'\n" );
 106                         $db = wfGetDB( DB_MASTER );
 107                         $db->delete( 'image',
 108                                 array( 'img_name' => $name ),
 109                                 __METHOD__ );
 110                 }
 111         }
 112
 113         private function filePath( $name ) {
 114                 if ( !isset( $this->repo ) ) {
 115                         $this->repo = RepoGroup::singleton()->getLocalRepo();
 116                 }
 117                 return $this->repo->getRootDirectory() . '/' . $this->repo->getHashPath( $name ) . $name;
 118         }
 119
 120         private function imageExists( $name, $db ) {
 121                 return $db->selectField( 'image', '1', array( 'img_name' => $name ), __METHOD__ );
 122         }
 123
 124         private function pageExists( $name, $db ) {
 125                 return $db->selectField(
 126                         'page',
 127                         '1',
 128                         array( 'page_namespace' => NS_FILE, 'page_title' => $name ),
 129                         __METHOD__
 130                 );
 131         }
 132
 133         private function pokeFile( $orig, $new ) {
 134                 $path = $this->filePath( $orig );
 135                 if ( !file_exists( $path ) ) {
 136                         $this->output( "missing file: $path\n" );
 137                         $this->killRow( $orig );
 138                         return;
 139                 }
 140
 141                 $db = wfGetDB( DB_MASTER );
 142
 143                 /*
 144                  * To prevent key collisions in the update() statements below,
 145                  * if the target title exists in the image table, or if both the
 146                  * original and target titles exist in the page table, append
 147                  * increasing version numbers until the target title exists in
 148                  * neither.  (See also bug 16916.)
 149                  */
 150                 $version = 0;
 151                 $final = $new;
 152                 $conflict = ( $this->imageExists( $final, $db ) ||
 153                                 ( $this->pageExists( $orig, $db ) && $this->pageExists( $final, $db ) ) );
 154
 155                 while ( $conflict ) {
 156                         $this->output( "Rename conflicts with '$final'...\n" );
 157                         $version++;
 158                         $final = $this->appendTitle( $new, "_$version" );
 159                         $conflict = ( $this->imageExists( $final, $db ) || $this->pageExists( $final, $db ) );
 160                 }
 161
 162                 $finalPath = $this->filePath( $final );
 163
 164                 if ( $this->dryrun ) {
 165                         $this->output( "DRY RUN: would rename $path to $finalPath\n" );
 166                 } else {
 167                         $this->output( "renaming $path to $finalPath\n" );
 168                         // @todo FIXME: Should this use File::move()?
 169                         $db->begin( __METHOD__ );
 170                         $db->update( 'image',
 171                                 array( 'img_name' => $final ),
 172                                 array( 'img_name' => $orig ),
 173                                 __METHOD__ );
 174                         $db->update( 'oldimage',
 175                                 array( 'oi_name' => $final ),
 176                                 array( 'oi_name' => $orig ),
 177                                 __METHOD__ );
 178                         $db->update( 'page',
 179                                 array( 'page_title' => $final ),
 180                                 array( 'page_title' => $orig, 'page_namespace' => NS_FILE ),
 181                                 __METHOD__ );
 182                         $dir = dirname( $finalPath );
 183                         if ( !file_exists( $dir ) ) {
 184                                 if ( !wfMkdirParents( $dir, null, __METHOD__ ) ) {
 185                                         $this->output( "RENAME FAILED, COULD NOT CREATE $dir" );
 186                                         $db->rollback( __METHOD__ );
 187                                         return;
 188                                 }
 189                         }
 190                         if ( rename( $path, $finalPath ) ) {
 191                                 $db->commit( __METHOD__ );
 192                         } else {
 193                                 $this->error( "RENAME FAILED" );
 194                                 $db->rollback( __METHOD__ );
 195                         }
 196                 }
 197         }
 198
 199         private function appendTitle( $name, $suffix ) {
 200                 return preg_replace( '/^(.*)(\..*?)$/',
 201                         "\\1$suffix\\2", $name );
 202         }
 203
 204         private function buildSafeTitle( $name ) {
 205                 $x = preg_replace_callback(
 206                         '/([^' . Title::legalChars() . ']|~)/',
 207                         array( $this, 'hexChar' ),
 208                         $name );
 209
 210                 $test = Title::makeTitleSafe( NS_FILE, $x );
 211                 if ( is_null( $test ) || $test->getDBkey() !== $x ) {
 212                         $this->error( "Unable to generate safe title from '$name', got '$x'" );
 213                         return false;
 214                 }
 215
 216                 return $x;
 217         }
 218 }
 219
 220 $maintClass = "ImageCleanup";
 221 require_once RUN_MAINTENANCE_IF_MAIN;